About the client
The largest Russian company specializing in air separation technologies and equipment, production of process gases and development of end-to-end solutions for bypass gas, natural gas, and LNG processing.
The company needed a service for remote monitoring of air separation plant and collection of data with a specific set of metrics. The project objective was development of a mathematical model for optimization of the process of technical maintenance of equipment and a number of related business processes, including: procurement, supply, and client relationship management, financial and production planning.
The service should collect, transform, and store data for further retrospective analysis, as well as for further solutions:
- for remote diagnostic maintenance (the service ensures real-time integration of telemetric data into the mathematic model for predicting the outage of air separation units and aggregates);
- for prognostic service (allows predicting the time of pump maintenance on remote pump stations to avoid unplanned equipment failures and stoppages);
- to determine the production function (searching for various combinations of production factors for maximum production output);
- to analyze the impact of temperature, humidity, and ambient air composition on performance.
The plant operation principle is as follows: ambient air passes through a number of mechanic filters, comes to the compressor for compression and to the liquefying section for further air separation into components.
The air separation unit has gas analyzers, flow meters, power metering units (sensors used within the frames of this project), and a SCADA cabinet with a controller and a field bus.
44 parameters to be tracked and logged have been defined within the frames of the pilot project. They include argon fraction, argon, nitrogen, oxygen, air incoming into the separation unit, and information from the power units of the air separation plan: active, reactive, and total energy.
A cloud service powered by Microsoft Azure.
At the first stage of the project, it was necessary to select a gateway that will act as middleware between the cloud and switching equipment of the air separation unit.
The client has chosen the ThingsPro Gateway based on UC-8112, a device certified by Microsoft. The manufacturer, MOXA, is a wide-known SCADA vendor, and the device selection caused no concern from the client.
From the hardware point of view, ThingsPro Gateway is an embedded computer for industrial applications. It can work in unfavorable conditions and use traditional RS-422/485 interfaces to connect to the PLC and I/O modules. Wire-based Ethernet or LTE are used for Internet access.
Software gateway has a convenient web interface that enables quick and simple data transfer into the cloud. The gateway sends a request to the controller and packs the project-related data that are then sent to Azure.
The IoT Hub streaming data processor, Stream analytics, SQL Database and Power BI have been configured in the cloud platform.
Azure application architecture
- IoT Hub accepts messages from the ThingsPro Gateway.
- Stream Analytics addresses the IoT Hub, takes the messages from there and processes them.
- The Stream Analytics stream is then branched: the first part goes directly to Power BI and real-time dashboards and the second part is sent to SQL Database, where there are around 160 tables. Stored procedures in SQL Database are executed according to a specific schedule or based on specific algorithms (incident aggregates and trackers). The goal of aggregates is to produce hourly values from second-to-second values and then transform hourly values into daily, monthly, quarterly, half-annual and annual values. The incident trackers have rated values for each parameter. When the values are in the allowed range, the plant operates normally. It also has emergency values: when received, they are registered in a separate table for further study by the plant process manager to understand the things that happened 10 minutes before the incident and 10 minutes after the incident.
- SQL Database is connected to Power BI for historical analysis. With a dashboard in Power BI, you can automatically move the results into the mobile version.
- After configuration of all logics in SQL Database, we have started Power BI configuration. We have created dashboards reflecting the process in real-time mode, and additional schedules for the key operation parameters, to which the borders are applied (rated boundary range and rated emergency range).
Furthermore, the following reports have been configured:
- shift-to-shift specific energy consumption (the production is very energy-intensive and energy is the main cost item);
- finished product output with predictions based on a one-factor model and built-in Power BI tools;
- power consumption for each air separation unit (analog of electric power technical metering system) and instant power metering;
- analysis of production incidents to study the transient processes.
The IoT service for remote monitoring of an air separation plant operation and data collection has been successfully commissioned. Implementation in other air separation plants has been planned.
- These services are especially suitable to companies that plan to change their mechanized equipment to semi-automated and fully automated units. Thanks to equipment controllers, it is possible to host data in storage and implement a part of MES functions.
- The payment for the service depends on the number of operations. Subscription can be cancelled.
- Recruiting analysts is much simpler than developers of industry-specific systems.
- The platform has additional services that can be aggregated with integration mechanisms built into the platform.
- Open access to technical and reference documentation helps building competences within the company.
- No need to worry about SLA: everything lies within the responsibility of a data center service provider.
Project Director at the Strategic Project Management Division, Softline